Background

Immunotherapy with immune checkpoint inhibitors has emerged as a novel option for many patients with advanced cancers whom previously had limited treatment options and experienced poor outcomes. While clinical studies have demonstrated survival benefits and durable responses in various cancer entities, immunotherapy of gynaecological cancers is still in relative infancy. In the United States, pembrolizumab, an antibody against programmed cell death protein 1 (PD-1), is approved for advanced endometrial cancers with high levels of microsatellite instability (MSI-high) and for recurrent or progressive metastatic cervical cancer positive for programmed death-ligand 1 (PD-L1), suggesting that these parameters could serve as predictive biomarkers.

Under physiological conditions, immune checkpoints play a crucial role in preventing autoimmunity [1]. Through the expression of PD-L1, cancer cells modulate the immune checkpoint to downregulate T cells, thereby protecting themselves from immune attack [2]. Immune checkpoint inhibitors reduce the interaction between cancer cells and T cells so that re-activated lymphocytes can destroy their malignant targets. Antibodies against PD-1 such as nivolumab have recently been approved for the treatment of different tumour types, particularly melanoma and non-small cell lung cancer [2].

More than 95% of cervical cancers are caused by a human papillomavirus (HPV) infection. Despite Papanicolaou testing, HPV screening, and prophylactic HPV vaccination, cervical cancer remains the fourth most common cause of death by cancer in women worldwide. The 5-year overall survival (OS) of recurrent or metastatic cervical cancer is poor (around 15%), mainly due to limited treatment options.

For the last 20 years, the standard of care for patients with recurrent or metastatic cervical cancer has been cisplatin-based chemotherapy combined with paclitaxel. However, despite the addition of other agents such as vinorelbine, gemcitabine or topotecan [3], most patients deteriorated rapidly, developing platinum resistance at recurrence [4]. Adding bevacizumab, an anti-vascular endothelial growth factor antibody, to chemotherapy became the standard of care since the Gynecologic Oncology Group 240 trial demonstrated a survival benefit of almost 4 months [5]. Recent evidence indicates the potential usefulness of immune checkpoint inhibitors in cervical cancer [4] with objective response rates (ORR) in recurrent and/or advanced cervical cancer ranging from 12.2 to 26% [6]. However, it remains challenging to select patients with a potential therapeutic response and to anticipate the magnitude of response. Predictive markers of therapeutic response to immunotherapy are still unclear. Increased PD-L1 expression by tumour and immune cells as well as elevated tumour mutational burden (TMB; mutations per tumour genome coding region) have been associated with an increased likelihood of response [7]. High-TMB tumours tend to have more neoantigens serving as potential targets for the immune system [8]. Furthermore, patients with microsatellite-unstable (MSI-high) tumours seem to qualify for immunotherapy regardless of tumour type [9]. Last, immune responsiveness signatures based on methylation profiles, as already identified in lung cancer, are rapidly moving into clinical focus [10].

Whereas immune checkpoints are crucial in maintaining self-tolerance, their inhibition by immunotherapy puts the immune system into a hyperactive state that can cause immune-related adverse events (IRAE). IRAEs are frequent and occur in up to 70% of patients treated with anti-PD1/PD-L1 antibodies. Typically, IRAEs occur within weeks to a few months after treatment initiation, yet, reports on delayed IRAEs one year after treatment discontinuation have also been recorded [11]. Any organs can be affected by an IRAE, which can, in some cases, be life-threatening or even lethal [11].

We report a patient with persistent complete response after third-line treatment with Nivolumab for relapsed chemotherapy-resistant, high-TMB cervical cancer, which underlines the potential of immunotherapy. In addition, we describe vulvitis as an isolated IRAE under nivolumab, which was successfully treated with topical corticosteroids.

Case presentation

A 62-year-old woman was diagnosed with squamous cell carcinoma of the cervix uteri FIGO Stage IIIC2(r) (according to FIGO 2018). Computed tomography (CT) at first diagnosis showed a large cervical mass of 9 cm, enlarged iliac and paraaortic lymph nodes, and right hydronephrosis. Prior personal and family history were negative. First-line treatment included combined radio-chemotherapy with 6 cycles of cisplatin 40 mg/m² weekly and lymph node irradiation, which was followed by intra-cervical brachytherapy with Iridium-192 resulting in a partial response. Three months later, progressive disease with a new palpable mass in a left supraclavicular lymph node and suspicion of pulmonary metastases (small nodules of 6 mm, not amenable to biopsy) was noted. A mixed response of the pelvic nodal metastases as well as a progression of the cervical mass was documented. After radiotherapy of the left supraclavicular lymph node metastasis and second-line chemotherapy with 6 cycles of carboplatin/AUC6, paclitaxel 175 mg/m², and bevacizumab 15 mg/kg, partial response (regression of all lesions without complete resolution) for 6 months was achieved. However, bevacizumab had to be interrupted due to rectorrhagia and ensuing anaemia. Due to progressive disease with new retrocrural and paraesophageal lymph node metastases, local radiotherapy was applied and a third-line systemic treatment with nivolumab (3 mg/kg q2w for 17 months, then according to Checkmate-358 with 240 mg q2w) was initiated leading to complete remission as verified by positron emission tomography-computed tomography (PET-CT) at 12 months after immunotherapy initiation. Complete remission was also documented after 22 months of treatment.

The immunotherapy was well tolerated over the course of 8 months, then the patient presented with a well-delimited inflammation of the vulva and perianal region with ulceration and epidermolysis (Fig. 1a). A bacterial, fungal, or viral infection was excluded by microbiological analyses. Histology of a punch biopsy revealed a lichenoid interface dermatitis with a pronounced lymphocytic infiltrate along the dermo-epidermal junction as well as intraepithelial. Immunohistochemically, most of lymphocytes were CD8-positive T cells intermixed with CD4-positive T cells (Fig. 2), consistent with an immunotherapy-related vulvitis. This IRAE was successfully treated by topical corticosteroid (clobetasol propionate 0.05%, 1-2x daily; Fig. 1b).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1

Clinical appearance and successful treatment of the IRAE. a Ulcerative vulvitis six months after initial IRAE symptoms; biopsy shown in Fig. 2 was taken. b Successful management of IRAE with topical corticosteroids

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2

a-c Biopsy of the vulva with lichenoid interface dermatitis: Squamous epithelium with hyper- and parakeratosis, irregular acanthosis, and dense band-like and perivascular lymphocytic infiltrate along the dermo-epidermal junction as well as intraepithelial lymphocytic infiltrate. Scattered necrotic keratinocytes (original magnification a 40x, b 100x, c, 200x). e-f Immunohistochemistry for T cell markers CD3, CD4, and CD8: Intraepithelial T cells predominantly express CD8 (original magnification e, f 100x)

We performed molecular analyses of the primary tumour. The tumour was positive for PD-L1 (clone SP263, Ventana), as 30% of tumour infiltrating immune cells were PD-L1 positive (defined as the proportion of tumour area occupied by PD-L1 staining of any intensity in immune cells), whereas the tumour cells were PD-L1 negative. Unexpectedly, the tumour was negative for p16 by immunohistochemistry, so that an HPV-association appeared unlikely. A qPCR-based HPV test (Anyplex™ II HPV28 Detection, Seegene) on DNA from formalin-fixed, paraffin-embedded tissue was negative. To further characterize the neoplasm, we performed a genome-wide DNA methylation analysis (Infinium Methylation EPIC, Illumina) and compared the methylation pattern by a combined filtering and dimension reduction approach [12, 13] with reference data on various cancer types (The Cancer Genome Atlas, TCGA) including HPV-associated cervical squamous cell carcinomas, as specific HPV-associated structural aberrations and increased target-gene expression have been reported [14]. We observed the highest concordance with the methylation classes of squamous cell carcinomas, particularly of the cervix, head and neck, and oesophagus (Fig. 3a. Methylation array data were also analysed for copy number changes, revealing a relatively flat profile (Fig. 3b,c).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3

Microarray data analysis: a Region of interest within UMAP plot of the top 25′000 differentially methylated sites in the genome reveals that the current case (magenta arrow) does not fall into the centre of classical cervical squamous cell carcinomas (yellow) but overlaps with the methylation classes of squamous cell carcinomas of various origins (other colors); brown dots at bottom: bladder cancer. b, c Copy number profiles of tumour-derived DNA, calculated with R/conumee from methylation array data; b this case; c typical copy number profile of cervical squamous cell carcinoma (from TCGA reference cohort)

TMB was exceptionally high with 44 mutations/megabase (Oncomine™ Tumor Mutation Load Assay, Thermo Fisher Scientific), even though cut-off levels for non-HPV-related cervical cancers have yet to be defined. The TMB sequencing data were further screened for disease-related somatic variants. They revealed a pathogenic mutation in the PIK3CA gene and a likely pathogenic mutation in the ERBB2 gene with allelic frequencies close to 40%, matching the estimated tumour proportion of 80%). Somatic mutations in both of these genes have been suggested to play a role in the pathogenesis of cervical squamous cell carcinoma [15].

Immunohistochemistry for DNA damage repair proteins (MLH1, MSH2, MSH6, PMS2) showed preserved expression of all examined proteins, consistent with a microsatellite stable (MSS) carcinoma.

Funding

Not applicable.